【摘要】 工件型面的精确修磨是现代制造最常见的一种技术,为了实现型面的精确修磨必须要准确采集和测量型面的几何信息并通过特征建模为精确打磨提供条件。针对激光扫描仪获取的工件表面三维点云信息,依据点云特征分布情况,提出了一种改进的随机采样一致性点云分割算法以获取工件型面特征,该算法首先利用八叉树(Octree)建立点云之间拓补几何关系,然后根据K邻域搜索的欧式距离判别对点云进行初始分类,将分类结果作为随机采样一致性(RANSAC)初始种子点选取区域。最后利用面片间法向量和欧式距离判别对分割结果进行优化。工程运用表明该方法能够有效获取工件型面特征,测量精度满足工程需要。更多还原

【Abstract】 Precise grinding of workpiece profiles is one of the most common techniques in modern manufacturing. In order to achieve accurate grinding of the profile,it is necessary to accurately acquire and measure the geometrical information of the profile and provide conditions for precise grinding through feature modeling. According to the 3 D point cloud information of the workpiece surface acquired by the laser scanner,an improved random sampling consistency point cloud segmentation algorithm is proposed to obtain the surface features of the workpiece according to the distribution of point cloud features.Firstly,the algorithm uses the Octree tree(Octree)to establish the topological geometric relationship among point clouds,and classifies the point clouds according to the Euclidean distance of the K-domain search. The classification results are selected as the random seed consistency(RANSAC)initial seed point area. The segmentation results are optimized by the inter-pane normal vector and the Euclidean distance discrimination. The engineering application shows that the method can effectively acquires the surface features of the workpiece,and the measurement accuracy meets the engineering needs.更多还原